Lazy document attribute

[marcosgz]

Lazy Document attributes

The idea is to define document attributes that can be lazy loaded and updated using update or bulk update operations. This is useful for example when a document has nested and denormalized fields that are expensive to update the whole document. So, you can retrieve the value of the attribute and partially update the document.

The lazy document attribute can be defined on the index repository level:

class PostsIndex < Esse::Index
  repository :post do
    colection do
      # ...
    end

    document do |post|
      {
        _id: post.id,
        title: post.title,
      }
    end

    lazy_document_attribute :comments do |posts|
      Comment.where(post_id: posts.map(&:id)).group(:post_id).transform_values(&:count)
    end
  end
end

Note that the result of lazy_document_attribute is a hash where the key is one of the given post or its id.

So you to partially update the document by updating the comments count:

PostsIndex::Post.update_documents_attribute(:comments, postid1, postid2)

This could be part of some callback in the model or a background job.

You can also import documents by eager loading the lazy attribute or lazy loading it:

# Eager loading it, means that the attribute will be added to the document before bulk indexing it
PostsIndex.import(context: context, eager_include_document_attributes: [:comments])

# Lazy loading it, means that the attribute will be add to the document with a bulk update request after bulk indexing it
PostsIndex.import(context: context, lazy_update_document_attributes: [:comments])